||Earthwork projects are performed in an environment fraught with uncertainty and using expensive machines that interact in complex ways. While the planning of these projects can be improved significantly with discrete-event simulation modeling, most projects are still planned using traditional tools. Recent advances in general-purpose simulation tools for construction operations modeling, such as STROBOSCOPE, enable accurate and detailed modeling of any complex operation. These tools, however, demand a level of training that is beyond that which can be found in most current construction practitioners. Special-purpose simulators bring the benefits of simulation modeling to users with little or no training by providing a modeling environment that is very close to the problem and geared towards a narrow domain. This paper presents EarthMover, a special-purpose simulation-modeling tool for planning and estimating earth-moving operations. The paper explores the use of EarthMover as a state-of-the-art simulation-modeling tool that allows construction practitioners to model complex earthwork operations in detail. EarthMover supports model definition via an interactive graphical interface where users drag and drop elements such as road segments, bridges, and loading and hauling units. In addition, EarthMover produces static and dynamic output. The static output includes numerous graphs and tables that describe the performance of the system in detail. The dynamic output includes an animation of the simulated system where the loading and hauling units can be seen moving about the site as they perform construction. The animation serves as a verification, validation, presentation, and credibility tool. The paper also presents the tools and techniques that were used to build EarthMover. These tools and techniques can be readily used to build other special-purpose simulators for building construction as well as for other types of heavy construction. The tools include STROBOSCOPE as a simulation engine, Visio for the graphical and interactive model definition, and Proof Animation for the dynamic output. The techniques include Visual Basic programming and OLE automation.